NatWest signs collaboration agreement with OpenAI

Congratulations to the team at NatWest - they've just announced that they've signed what I imagine is a super-charged Enterprise-style agreement with OpenAI.

This is an important step for any financial services player that's trying to do anything material when it comes to leveraging tools from players like OpenAI.

There are usually what feels like a billion different hurdles to getting anything done in financial services when it comes to anything related to technology and more specifically with AI. And those hurdles need to be surmounted before you can get to the comparative luxury of worrying about AI model committees and so on.

Much of the base-level challenge for most banks tends to come down to difficult but important policy issues around data control, information security and so on.

I wrote recently of OpenAI's ChatGPT announcement about the creation of a European data residency offering for enterprise clients.

Quite a lot of companies have had to either ignore, soften or explicitly 'handle' (a deliberately nebulous term) the standard data residency and transmission protocols they typically demand in the rush to get access to models and capability from many US-based AI vendors.

The first step for any big company is an enterprise agreement with the vendor. This gives them the ability to start to dictate terms – or, at least, discuss the issues surrounding them and begin to address all of the compliance, security and other related challenges. A lot of lawyer time – actual humans (at least, on the part of NatWest) will have been spent making sure that they've explicitly defined all of the regulatory expectations (and more) surrounding the use of OpenAI tooling. In some cases it might have needed OpenAI to agree to adjust policies or implement some specific technical demands. In other cases, it's about making sure that – if the absolute worst happens – the bank at least tried to do the right thing.

But now comes the exciting bit.

For a chap like me, there is something absolutely astonishingly exciting about having unrestricted usage of technology in the enterprise that you're trying to change. Having worked in Transformation for the longest time, I very much appreciate the value and opportunity presented by enterprise level agreements. Most big companies, for example, will have something similar with Microsoft, IBM or others. Essentially it means that having properly engaged with each other and constructed an acceptable agreement between both parties, you can now get on with driving value.

Although there are quite a few exceptions, the fundamental point of signing an enterprise agreement with Microsoft means that, essentially, you can do-what-you-want, because you've explicitly managed all of the normal issues that prevent you from working seamlessly with vendors (everything from data management to security protocols). And the money bit is sorted too.

This means that if, for example, Microsoft releases a new app, as long as it's covered by their enterprise terms, it'll sit in our existing tenancy with all of the security and assurances already in place with the stuff we're already using. So it's often just a one page variation agreement needed for your company to take advantage of the technology. In some cases you can simply just press a button on an administrator console and the new tech is deployed! This why, for example, a load of companies adopted Microsoft Teams rather than Slack. Teams was point-and-click to deploy. Slack needed a ton of engagement with procurement, Legal, IT Security, Compliance... blah blah blah.

So this is the exciting bit. This is what NatWest is looking at.

It's why people like Miles Hillier (Product & Technology Director, AI, NatWest - and a super chap) are waxing lyrical on LinkedIn today about the OpenAI agreement.

After a heck of a lot of work, they've now got access to a huge toy box from OpenAI. Everything from Agents, Whisper, ChatGPT obviously and ChatGPT's super-duper $200/month Deep Research and so on. That's not all. They've got access to the internal OpenAI roadmap (with some restrictions I'm sure) and the ability to start road-testing OpenAI's gizmos before they hit the market.

NatWest has now 1000x'd its AI capabilities

This, dear reader, is market-moving. I'd suggest this is market-moving news. It indicates – or at least suggests – that NatWest has now 1000x'd its AI capabilities when compared to its position this time yesterday.

Now it's worth pointing out that NatWest has hardly been a laggard when it comes to AI. I think you can fairly argue that the bank has been leading the way in the United Kingdom with its efforts to test, learn, play with and harness AI technology.

No wonder Miles and the team are exuberant.

If you're sitting at one of the other European banks that don't have an enterprise agreement with OpenAI, Anthropic or similar, then you're likely to be left behind in the dust. I can name a few banks whose leaders (even today!) are actively – and I do mean actively – preaching NO material use of AI, at all, throughout their institutions citing hallucination as a primary barrier.

The learning, the awareness and the experience (if they do it right) that NatWest will get from this partnership can be enormously powerful to their business.

Imagine being able to use ChatGPT properly in your job. Actually properly. No elicit 0cutting and pasting. Imagine being able to put some of OpenAI's agents to work processing a lot of the grunt work your teams usually have to do. Imagine being able to get Deep Research to go and look at every single piece of data in your estate – as well as the total of all human knowledge.

So nice work Team NatWest. I'm looking forward to seeing what you do with it!

For everyone else working in Financial Services wanting to be in a similar position to NatWest, send this article and the NatWest press release to your boss.